- 學位論文
使用光聲效應量測葡萄糖濃度之效能探討
Performance Issues in Glucose Concentration Measurements Using Photoacoustics
摘要
本論文主要探討使用超音波聲速與光聲振幅變化量測葡萄糖濃度之可行性。對於糖尿病患而言，血糖量測是必要的檢測。目前最普遍的方法仍以侵入式的化學檢測法為主，而這會造成受測者傷口及痛苦。利用光聲效應最大的潛在優點為非侵入式及同時利用物質的光學特性以及聲學特性來判斷葡萄糖的濃度，本論文之研究目標即為探討此可行性。在光聲實驗架構方面，主要以Ti:sapphire作為雷射之光源，發出波長在940 nm的近紅外光照射葡萄糖溶液，並且用中心頻率為1 MHz，聚焦深度為1.27 cm之線聚焦超音波探頭以側向模式進行光聲波量測，並將所得訊號之振幅差異來分辨葡萄糖之濃度。為降低雷射輸出之變異所造成之影響，亦利用同步量測之雷射能量做修正。此結果顯示在光聲訊號的平均振幅方面，濃度每增加1%，平均振幅可上升2%，此與參考文獻數據相符合。光聲訊號振幅的標準差則大約是振幅的1.5%-2%，此誤差限制了實際應用之可行性。目前所知誤差來源包括系統內部所產生的熱雜訊，此經由訊號平均可有效降低；誤差來源也包括了雷射能量的變異，此則無法被完全去除。在聲速量測方面，使用中心頻率為20 MHz之探頭，在距離探頭下1公分處放置鐵板，量測脈衝回波。在距離為1公分之下，葡萄糖濃度每上升1%，飛行時間可下降約0.18%左右，而標準差大約是0.2 ns，其主要來自於jitter之影響。本研究目前結果仍以溶液實驗為主，此系統可分辨之濃度差為1%，而美國食品藥物管理局(FDA)則規定需能分辨濃度差為0.01%才可供人體使用，因此需要再降低標準差，並結合聲速參數去除受外界干擾之訊號。本研究也進行了血紅素溶液之初步實驗。在光聲訊號的平均振幅方面，葡萄糖濃度每上升1%，平均振幅會上升1.4%。在量測聲速方面，與純水溶液所得結果相似，聲速會隨葡萄糖濃度上升，而葡萄糖溶液濃度每上升1%，飛行時間下降約0.17%。In vivo的實驗將在未來進行，由於紅血球會造成光的散射，而且血漿中其他物質和葡萄糖吸收光譜重疊的問題，將不易產生良好的光聲訊號，因此可探討以多波長的方式來逐一消除其他物質的干擾及選擇適當量測部位來克服in vivo應用之相關問題。
並列摘要
In this research, glucose concentration measurements based on photoacoustic(PA) signal amplitude and acoustic velocity are studied. Determination of blood glucose level is a required procedure in diabetes care. Currently, the most common method involves collecting blood samples for chemical analysis, but it is invasive and prone to result in pain and skin injury. It is the purpose of this study to investigate the feasibility of non-invasive glucose measurements based on PA and ultrasonic measurements. The PA setup includes a source Ti:sapphire laser and a line-focused single-crystal ultrasound transducer operating at 1 MHz. The transducer was arranged perpendicular to the incident laser beam for sideward detection. Our experiment results reveal that the PA amplitude increases by about 2% when the glucose concentration increases by 1%, and the standard deviation is around 1.5% to 2%. The standard deviation results from thermal noise, which can be decreased by signal averaging and the instable laser output. On the other hand, sound velocity was estimated by pulse echo measurements. The experimental setup included an ultrasound transducer operating at 20 MHz, with the distance between the ultrasound transducer and a reflector being 1 cm. The experiment results show that the time of flight decreases by 0.18% when the glucose concentration increases by 1% and the standard deviation is around 0.2ns, which is mainly caused by jitter. In conclusion, our measurements show that a 1% concentration change is detectable. However, the FDA requires a 0.01% detectability. We have also performed preliminary experiments with hemoglobin solution. Results show that the PA amplitude increases by about 1.4% and the time of flight decreases by about 0.18% when the glucose concentration increases by 1%. Future works will focus on improving the method by using multiple-laser wavelength, and to study influence of blood constituents.